Establishment of an in‐house real‐time RT‐PCR assay for the detection of severe acute respiratory syndrome coronavirus 2 using the first World Health Organization international standard in a resource‐limited country

Abstract Background The COVID‐19 pandemic caused by SARS‐CoV‐2 remains public health burdens and many unresolved issues worldwide. Molecular assays based on real‐time RT‐PCR are critical for the detection of SARS‐CoV‐2 in clinical specimens from patients suspected of COVID‐19. Objective We aimed to establish and validate an in‐house real‐time RT‐PCR for the detection of SARS‐CoV‐2. Methodology Primers and probes sets in our in‐house real‐time RT‐PCR assay were designed in conserved regions of the N and E target genes. Optimized multiplex real‐time RT‐PCR assay was validated using the first WHO International Standard (NIBSC code: 20/146) and evaluated clinical performance. Results The limit of detection validated using the first WHO International Standard was 159 IU/ml for both E and N target genes. The evaluation of clinical performance on 170 clinical samples showed a positive percent agreement of 100% and the negative percent agreement of 99.08% for both target genes. The Kappa value of 0.99 was an excellent agreement, the strong correlation of C t values observed between two tests with r 2 = 0.84 for the E gene and 0.87 for the N gene. Notably, we assessed on 60 paired saliva and nasopharyngeal samples. The overall agreement was 91.66%, and Kappa value of 0.74 showed a high agreement between two types of samples. When using nasopharyngeal swabs as the reference standard, positive percent agreement, and negative percent agreement were 91.83% and 90.90%, respectively. Conclusion In the present study, we established and validated an in‐house real‐time RT‐PCR for molecular detection of SARS‐CoV‐2 in a resource‐limited country.


| INTRODUC TI ON
The outbreak of coronavirus disease 2019 (COVID-19) pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) remains a big challenge and public health burden in many countries associated with morbidity and mortality. According to the World Health Organization (WHO) report, COVID-19 case incidence is still increasing in many regions on the world. Globally, there were nearly 265 million cases confirmed with SARS-CoV-2 infection and There were many commercial and laboratory-developed real-time RT-PCR assays have become available for routine diagnosis of SARS-CoV-2 in clinical laboratories. However, some real-time RT-PCR kits failed for the detection of SARS-CoV-2 associated with the presence of mutations in primer and probe binding regions. [8][9][10] Therefore, this study aimed to establish and validate an in-house real-time RT-PCR for the detection of SARS-CoV-2 in resource-limited settings.

| Primers and probes design
The complete sequences of SARS-CoV-2 were downloaded from the Genbank and GISAID database for alignment using BioEdit 7.0 (https:// bioed it.softw are.infor mer.com/versi ons/) to select highly conserved regions of targets in the E and N genes for design primers and probes with the aid of Primer Express software version 3.0 (Thermo Fisher Scientific). The primers and probe set of N target are specific to SARS-CoV-2, whereas E gene primers and probe will detect both SARS-CoV and SARS-CoV-2 but not detect other coronaviruses. In addition, we used primers and probe set of Rase P published by USA-CDC as an internal control in our assay. 11 Primers and probes sequences are summarized in Table 1. Primers and probes were purchased from IDT.

| Real-time RT-PCR:
We optimized a multiplex real-time RT-PCR assay (designated as Laboratory Developed Assay: LDA assay) in a total volume of 20 µl containing 10 µl of 2X Luna ® Universal probe one-step RT-qPCR Kit (New England Biolab), 1 µl of RT enzyme, 1 µl each of 10 µM forward and reverse primers of N and E genes, and the 0.4 µl each of 5 µM probes of N and E genes, 0.8 µl each of 10 µM forward and reverse primers of Rnase P and 0.4 µl of 5 µM Rnase P probes and 5 µl of RNA template.
All reactions were run on a RotorGene Q 5plex MDx (Qiagen) using the following thermal cycling conditions: 50°C for 2 min, followed by 45 cycles of 90° C for 15 min, 94° C for 15 s, and 58°C for 60 s.

| LiliF COVID-19 real-time PCR kit
LiliF COVID-19 real-time RT-PCR (iNtRON Biotechnology) used in this study as the reference assay to validate our established in-house LDA assay on clinical samples. This kit was designed for the detection of SARS-CoV-2 using three target genes: envelope (E), RNAdependent RNA polymerase (RdRp), nucleocapsid (N), and Rnase P as internal control. According to the manufacturer's interpretation, diagnosis of SARS-CoV-2 is confirmed for a sample that has at least three genes with cycle threshold (C t )-value ≤35.

| Validation of real-time RT-PCR using the first WHO International standard
The first WHO International Standard for SARS-CoV-2 RNA for Nucleic acid Amplification Technique (NAT)-based assays (NIBSC code: 20/146) was provided as a kind gift from Dr. Do Minh Si, Nanogen Biopharma, Vietnam. This material was reconstituted in 0.5 ml of molecular grade water to obtain the final concentration of K E Y W O R D S clinical performance, real-time RT-PCR, SARS-CoV-2

| Clinical specimens
Nasopharyngeal specimens were collected from patients sus-

| Sequencing confirmation and phylogenetic analysis
To assess the ability of the proposed method to amplify the SARS-CoV-2 variants. A 725 bp fragment of Spike gene was amplified by one-step RT-PCR (Qiagen GmbH) from samples detected positive by our method using primer sequences previously described. 13 In addition, we designed primers to amplify a 438-bp fragment covering primers and probe sequence region of N target gene. Amplified fragments were visualized under UV light, and then purified and sequenced using a 3130 XL sequencer. Sequences obtained were aligned with reference sequences retrieved from GenBank and GISAID ( Table 7) using Bioedit 7.0 (https://bioed it.softw are.infor mer.com/versi ons/) and MEGA 7.0 software (www.megas oftwa re.net). Phylogenetic tree was constructed using the neighborjoining method, and significance level was estimated with 1000 bootstrap replicates.

| Statistical analyses
Statistical analyses were done with SPSS 20.0 (IBM). The diagnostic agreements were analyzed to estimate confidence intervals (95% CI) for positive percent agreement (PPA) and negative percent agreement (NPA). Cohen's kappa values were calculated for evaluating overall agreement and comparing assays. Correlation analysis was used to evaluate the C t values of the positive results. p values <0.05 were considered as statistically significant.

| Analytical sensitivity
A serial dilution of standards as prepared above and real-time RT-PCR reactions were run in eight replicates of each dilution. Results were presented Table 2. By using probit analysis, the LOD determined for both targets were 159 IU/ml.
The reproducibility of real-time RT-PCR assay was assessed by coefficient of variation (CV) of the cycle threshold (C t ) values in the intra-and inter-assays at three different concentrations using the first WHO International Standard (5.02 × 10 5 IU/ml, 5.02 × 10 4 IU/ ml, 5.02 × 10 3 IU/ml). For intra-assay repeatability, each concentration was tested triplicate in one reaction, for inter-assay reproducibility, each concentration was run in three independent reactions across three different days. The results are presented in Table 3. The mean CV of C t values was observed to be lesser 5% for both concentrations evaluated, found an accurate and a good repeatability of our LDA assay.  While the NPA was 99.08% for both target genes ( Table 4) The overall agreement between two types of samples was 91.66% (55/60). When using nasopharyngeal swabs as the reference standard, positive percent agreement and NPA were 91.83% and 90.9% for both target genes, respectively. There was a high agreement (Kappa value = 0.74) between the two types of specimens (Table 5).
However, the mean C t values of nasopharyngeal swabs were significantly lower than that of saliva for both E and N target genes.
Detailed data were presented in Table 6. This finding may be using  To avoid underdiagnoses or false-negative results due to single nu-   17 Interestingly, among positive samples, a strong correlation between the C t values of E and N genes observed between our LDA and LiliF (Figures 1 and 2). Our findings were similar with previous reports as LDA assays were compared with commercial molecular tests or WHO's protocol and modified CDC's panel. 25,26 Interestingly, no mismatches observed in primers and probes binding regions as compared to reference sequences of SARS-CoV-2 variants of concern used this study. Sequencing data of positive samples confirmed the accuracy of our LDA assay.
Further analysis, we evaluated on paired saliva-nasopharyngeal swab showed a significant agreement between two types of samples and using nasopharyngeal swab as the gold standard indicated a high PPA of 91.83% and the significant difference observed as compared mean C t values of two sample types ( Table 6). These study findings showed the similar sensitivity of saliva-based real-time RT-PCR compared with paired positive nasopharyngeal (NP) samples ranging from 84% to 100% has been reported by other authors. 27 Notably, these saliva specimens may be self-collection or outside of hospital setting without assisting of nurse and saliva collection F I G U R E 3 Phylogenetic tree was constructed using the Spike partial gene sequences (22,522)  This study result highlighted the benefit of an in-house real-time RT-PCR assay was standardized and validated with the WHO reference standards in response to urgent testing capacity and situation of global shortage of supply chain.
In conclusion, the present study, we developed and evaluated an affordable in-house real-time RT-PCR assay for the detection of SARS-CoV-2 in clinical specimens in patients suspected of COVID-19. This helps to benefit an affordable LDA assay for the effective control of COVID-19 in resource-limited settings.

CONFLICT OF INTEREST
The authors have no conflicts of interest to disclose.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data in this study can be obtained directly from the corresponding author.